Re-Thinking Internet Architecture Today’s Internet Original Design Goal, Philosophy and Principles End-to-End Principle and “Hourglass” Architecture of Internet Pros and Cons; Challenging Issues What have changed? What may have yet to come? Overlay Networks Future Internet Architectures? What are key challenges/issues? E.g., mobility, security, “services-oriented” … Diversity of “end systems”: laptops, cell phones, sensors, …

Network Architecture What is (Network) Architecture? not the implementation itself “design blueprint” on how to “organize” implementations what interfaces are supported where functionality is implemented Two Basic Architectural Principles Modularity (e.g., layering) how to break network functionality into modules End-to-End Argument where to implement functionality

Architectural Principles (not unique to networks!) Zhi-Li’s version (synthesized from various sources) End-to-end argument functionality placement Modularity Increase inter-operability and manage complexity vertical partition -> layered architecture horizontal partition? Keep it simple, stupid (KISS principle) Occam’s Razor: choose simplest among many solutions! complicated design increases system coupling (inter-dependence), amplifies errors, .. don’t over-optimize! Separating policies from mechanisms decouple control from data “semantics-free” Design for scale hierarchy, aggregation, …

Some Design/Implementation Principles virtualization indirection soft state vs. hard state fate sharing randomization expose faults, don’t suppress/ignore caching ……

Original Internet Design Goals [Clark’88] In order of importance: Connect existing networks initially ARPANET and ARPA packet radio network Survivability ensure communication service even with network and router failures Support multiple types of services Must accommodate a variety of networks Allow distributed management Allow host attachment with a low level of effort Be cost effective Allow resource accountability

Priorities The effects of the order of items in that list are still felt today E.g., resource accounting is a hard, current research topic Different ordering of priorities would make a different architecture! How well has today’s Internet satisfied these goals? Let’s look at them in detail

Summary: Internet Architecture TCP UDP Packet-switched datagram network IP is the “compatibility layer” Hourglass architecture All hosts and routers run IP Stateless architecture No per flow state inside network IP Satellite Ethernet ATM

Summary: Minimalist Approach Dumb network IP provide minimal functionalities to support connectivity Addressing, forwarding, routing Smart end system Transport layer or application performs more sophisticated functionalities Flow control, error control, congestion control Advantages Accommodate heterogeneous technologies (Ethernet, modem, satellite, wireless) Support diverse applications (telnet, ftp, Web, X windows) Decentralized network administration Beginning to show age Unclear what the solution will be  probably IPv6?

Questions What priority order would a commercial design have? What would a commercially invented Internet look like? What goals are missing from this list? Which goals led to the success of the Internet?

Requirements for Today’s Internet Some key requirements (“-ities”) Availability and reliability “Always on”, fault-tolerant, fast recovery from failures, … Quality-of-service (QoS) for applications fast response time, adequate quality for VoIP, IPTV, etc. Scalability millions or more of users, devices, … Mobility untethered access, mobile users, devices, … Security (and Privacy?) protect against malicious attacks, accountability of user actions? Manageability configure, operate and manage networks trouble-shooting network problems Flexibility, Extensibility, Evolvability, ……? ease of new service creation and deployment? evolvable to meet future needs?

End System Based Overlay/P2P Services/Solutions General Concept of Overlays Some Examples End-System Multicast Rationale How to construct “self-organizing” overlay Performance in support conferencing applications Internet Indirection Infrastructure (i3) Motivation and Basic ideas Implementation Overview Applications

Overlay Networks

Overlay Networks Focus at the application level

Overlay Networks A logical network built on top of a physical network Overlay links are tunnels through the underlying network Many logical networks may coexist at once Over the same underlying network And providing its own particular service Nodes are often end hosts Acting as intermediate nodes that forward traffic Providing a service, such as access to files Who controls the nodes providing service? The party providing the service (e.g., Akamai) Distributed collection of end users (e.g., peer-to-peer)

Routing Overlays Alternative routing strategies No application-level processing at the overlay nodes Packet-delivery service with new routing strategies Incremental enhancements to IP IPv6 Multicast Mobility Security Revisiting where a function belongs End-system multicast: multicast distribution by end hosts Customized path selection Resilient Overlay Networks: robust packet delivery

IP Tunneling IP tunnel is a virtual point-to-point link Illusion of a direct link between two separated nodes Encapsulation of the packet inside an IP datagram Node B sends a packet to node E … containing another packet as the payload A B E F Logical view: tunnel A B E F Physical view:

6Bone: Deploying IPv6 over IP4 A B E F IPv6 tunnel Logical view: Physical view: C D IPv4 Flow: X Src: A Dest: F data Src:B Dest: E A-to-B: E-to-F: B-to-C: IPv6 inside

MBone: IP Multicast Multicast IP multicast Delivering the same data to many receivers Avoiding sending the same data many times IP multicast Special addressing, forwarding, and routing schemes Not widely deployed, so MBone tunneled between nodes unicast multicast

End-System Multicast IP multicast still is not widely deployed Technical and business challenges Should multicast be a network-layer service? Multicast tree of end hosts Allow end hosts to form their own multicast tree Hosts receiving the data help forward to others

RON: Resilient Overlay Networks Premise: by building application overlay network, can increase performance and reliability of routing application-layer router Princeton Yale Berkeley Two-hop (application-level) Berkeley-to-Princeton route

RON Can Outperform IP Routing IP routing does not adapt to congestion But RON can reroute when the direct path is congested IP routing is sometimes slow to converge But RON can quickly direct traffic through intermediary IP routing depends on AS routing policies But RON may pick paths that circumvent policies Then again, RON has its own overheads Packets go in and out at intermediate nodes Performance degradation, load on hosts, and financial cost Probing overhead to monitor the virtual links Limits RON to deployments with a small number of nodes

Secure Communication Over Insecure Links Encrypt packets at entry and decrypt at exit Eavesdropper cannot snoop the data … or determine the real source and destination

Communicating With Mobile Users A mobile user changes locations frequently So, the IP address of the machine changes often The user wants applications to continue running So, the change in IP address needs to be hidden Solution: fixed gateway forwards packets Gateway has a fixed IP address … and keeps track of the mobile’s address changes gateway www.cnn.com

Unicast Emulation of Multicast End Systems Routers Gatech CMU Stanford Berkeley

Routers with multicast support IP Multicast Gatech Stanford CMU Berkeley Routers with multicast support No duplicate packets Highly efficient bandwidth usage Key Architectural Decision: Add support for multicast in IP layer

Key Concerns with IP Multicast Scalability with number of groups Routers maintain per-group state Analogous to per-flow state for QoS guarantees Aggregation of multicast addresses is complicated Supporting higher level functionality is difficult IP Multicast: best-effort multi-point delivery service End systems responsible for handling higher level functionality Reliability and congestion control for IP Multicast complicated Deployment is difficult and slow ISP’s reluctant to turn on IP Multicast

End System Multicast Overlay Tree Gatech Stan1 Stanford Stan2 Berk1 CMU Gatech Stan1 Stanford Stan2 Berk1 CMU Stan1 Stan2 Berk2 Overlay Tree Gatech Berk1 Berkeley

Potential Benefits Scalability Easier to deploy Routers do not maintain per-group state End systems do, but they participate in very few groups Easier to deploy Potentially simplifies support for higher level functionality Leverage computation and storage of end systems For example, for buffering packets, transcoding, ACK aggregation Leverage solutions for unicast congestion control and reliability

Design Questions Is End System Multicast Feasible? Target applications with small and sparse groups How to Build Efficient Application-Layer Multicast “Tree” or Overlay Network? Narada: A distributed protocol for constructing efficient overlay trees among end systems Simulation and Internet evaluation results to demonstrate that Narada can achieve good performance

Performance Concerns Delay from CMU to Berk1 increases Gatech Stan1 Stan2 Berk2 Gatech Berk1 Delay from CMU to Berk1 increases CMU Gatech Stan1 Stan2 Berk1 Berk2 Duplicate Packets: Bandwidth Wastage

What is an efficient overlay tree? The delay between the source and receivers is small Ideally, The number of redundant packets on any physical link is low Heuristic used: Every member in the tree has a small degree Degree chosen to reflect bandwidth of connection to Internet CMU CMU CMU Stan2 Stan2 Stan2 Stan1 Stan1 Stan1 Berk1 Gatech Berk1 Gatech Berk1 Gatech Berk2 Berk2 Berk2 High latency High degree (unicast) “Efficient” overlay

Why is self-organization hard? Dynamic changes in group membership Members may join and leave dynamically Members may die Limited knowledge of network conditions Members do not know delay to each other when they join Members probe each other to learn network related information Overlay must self-improve as more information available Dynamic changes in network conditions Delay between members may vary over time due to congestion

Performance Metrics Delay between members using Narada Stress, defined as the number of identical copies of a packet that traverse a physical link Berk2 CMU Stan1 Stan2 Gatech Berk1 Delay from CMU to Berk1 increases Stan1 Stress = 2 CMU Stan2 Berk1 Gatech Berk2

ESM Conclusions Proposed in 1989, IP Multicast is not yet widely deployed Per-group state, control state complexity and scaling concerns Difficult to support higher layer functionality Difficult to deploy, and get ISP’s to turn on IP Multicast Is IP the right layer for supporting multicast functionality? For small-sized groups, an end-system overlay approach is feasible has a low performance penalty compared to IP Multicast has the potential to simplify support for higher layer functionality allows for application-specific customizations

Internet Indirection Infrastructure (i3) Motivations Today’s Internet is built around a unicast point-to-point communication abstraction: Send packet “p” from host “A” to host “B” This abstraction allows Internet to be highly scalable and efficient, but… … not appropriate for applications that require other communications primitives: Multicast Anycast Mobility … Today’s internet is build around a point-to-point communication abstractions. While this simple abstraction allows Internet to be highly scalable and Efficient, it is not appropriate for application that requires other communication primitives such as multicast, anycast, mobility, and so on.

Why? Point-to-point communication  implicitly assumes there is one sender and one receiver, and that they are placed at fixed and well-known locations E.g., a host identified by the IP address 128.32.xxx.xxx is located in Berkeley This is because there is a fundamental mismatch between point-to-point communication abstraction and these primitives. In particulr, the point-to-point communication abstraction implicitly assumes that there is only one sender and on receivers an that they are placed at fixed and well-known locations. Multicast, anycast, and mobility violate at least one of these assumptions. With mobility end-hosts do not have fixed locations, with multicast there are more than one receiver and sender.

IP Solutions Extend IP to support new communication primitives, e.g., Mobile IP IP multicast IP anycast Disadvantages: Difficult to implement while maintaining Internet’s scalability (e.g., multicast) Require community wide consensus -- hard to achieve in practice During the last decade a plethora of solutions have been proposed to implement anycast, multicast, and mobility functionalities. These solutions can be classified into two catgeories: IP solutions and application level solutions. Example of IP solutions are mobile IP, IP multicast, and IP anycast Unfortunately, despite years of research these solutions have yet to be deployed on a large scale. There are many reasons for this. Two of them are

Application Level Solutions Implement the required functionality at the application level, e.g., Application level multicast (e.g., Narada, Overcast, Scattercast…) Application level mobility Disadvantages: Efficiency hard to achieve Redundancy: each application implements the same functionality over and over again No synergy: each application implements usually only one service; services hard to combine To get around these problems recently several application level solutions have been proposed… However, they have several disadvantages. First, it is hard to achieve efficiency…. That is, proposal for application level multicast don’t address mobility and vice-versa. Usually, a new abstraction require to deploy a new overlay.

Key Observation “Any problem in computer science can Virtually all previous proposals use indirection, e.g., Physical indirection point  mobile IP Logical indirection point  IP multicast “Any problem in computer science can be solved by adding a layer of indirection”

Build an efficient indirection layer i3 Solution Build an efficient indirection layer on top of IP Use an overlay network to implement this layer Incrementally deployable; don’t need to change IP IP TCP/UDP Application Indir. layer

Internet Indirection Infrastructure (i3): Basic Ideas Each packet is associated an identifier id To receive a packet with identifier id, receiver R maintains a trigger (id, R) into the overlay network id data id data Sender id R trigger Receiver (R) R data

Service Model API Best-effort service model (like IP) sendPacket(p); insertTrigger(t); removeTrigger(t) // optional Best-effort service model (like IP) Triggers periodically refreshed by end-hosts ID length: 256 bits

Mobility Host just needs to update its trigger as it moves from one subnet to another Receiver (R1) Sender Such an indirection layer would support a large number of services. For example, to achieve mobility an end host needs only to update its trigger with the new address when it moves from one Subnet to another. id R2 id R1 Receiver (R2)

Multicast Receivers insert triggers with same identifier Can dynamically switch between multicast and unicast id data id data R1 data id R1 Receiver (R1) Sender Multicast is strightforward to achieve. The only difference between multicast and unicast is that in the case of the multicast there are more than on hosts inserting triggers with the same ID id R2 R2 data Receiver (R2)

Anycast Use longest prefix matching instead of exact matching Prefix p: anycast group identifier Suffix si: encode application semantics, e.g., location Sender Receiver (R1) p|s1 R1 Receiver (R2) p|s2 R2 p|s3 R3 Receiver (R3) data p|a

Service Composition: Sender Initiated Use a stack of IDs to encode sequence of operations to be performed on data path Advantages Don’t need to configure path Load balancing and robustness easy to achieve Transcoder (T) idT,id data idT,id data id data Receiver (R) R data Sender T,id data id R idT T

Service Composition: Receiver Initiated Receiver can also specify the operations to be performed on data Firewall (F) R data id data id data F,R data Finally, IL supports composable services, I.e., performing on the fly transformation such as transcoding on the data packets as they travel through the network. To achieve this we replace the packet ID with a stack of Ids, where each identifier excepting the last one identifies a transformation to be aplied on packets. The advantage of this solution versus previously proposed solutions is that you don’t need to find and configure the path,(you just insert the Ids in the proper order). Load balancing and robustness are easy to achieve. Just have more servers implementing the same operations. If one fails, the other one will take transparently over. Receiver (R) Sender idF F idF,R data id idF,R

Quick Implementation Overview i3 is implemented on top of Chord But can easily use CAN, Pastry, Tapestry, etc Each trigger t = (id, R) is stored on the node responsible for id Use Chord recursive routing to find best matching trigger for packet p = (id, data)

Routing Example R inserts trigger t = (37, R); S sends packet p = (37, data) An end-host needs to know only one i3 node to use i3 E.g., S knows node 3, R knows node 35 3 7 20 35 41 37 R S trigger(37,R) send(37, data) send(R, data) Chord circle 2m-1 [8..20] [4..7] [21..35] [36..41] [40..3]

Optimization #1: Path Length Sender/receiver caches i3 node mapping a specific ID Subsequent packets are sent via one i3 node [8..20] [4..7] [42..3] 37 data [21..35] Sender (S) cache node [36..41] R data 37 R Receiver (R)

Optimization #2: Triangular Routing Use well-known trigger for initial rendezvous Exchange a pair of (private) triggers well-located Use private triggers to send data traffic [8..20] [4..7] 30 data 37 [2] [42..3] S [30] 2 S [21..35] Sender (S) [36..41] R data 2 [30] R [2] 30 R 37 R Receiver (R)

Example 1: Heterogeneous Multicast Sender not aware of transformations Receiver R1 (JPEG) id_MPEG/JPEG S_MPEG/JPEG id (id_MPEG/JPEG, R1) send(id, data) Sender (MPEG) send((id_MPEG/JPEG, R1), data) send(R1, data) R2 Receiver R2 send(R2, data) Finally, IL supports composable services, I.e., performing on the fly transformation such as transcoding on the data packets as they travel through the network. To achieve this we replace the packet ID with a stack of Ids, where each identifier excepting the last one identifies a transformation to be aplied on packets. The advantage of this solution versus previously proposed solutions is that you don’t need to find and configure the path,(you just insert the Ids in the proper order). Load balancing and robustness are easy to achieve. Just have more servers implementing the same operations. If one fails, the other one will take transparently over.

Example 2: Scalable Multicast i3 doesn’t provide direct support for scalable multicast Triggers with same identifier are mapped onto the same i3 node Solution: have end-hosts build an hierarchy of trigger of bounded degree R2 R1 R4 R3 g x (g, data) (x, data)

Example 2: Scalable Multicast (Discussion) Unlike IP multicast, i3 Implement only small scale replication  allow infrastructure to remain simple, robust, and scalable Gives end-hosts control on routing  enable end-hosts to Achieve scalability, and Optimize tree construction to match their needs, e.g., delay, bandwidth

Example 3: Load Balancing Servers insert triggers with IDs that have random suffixes Clients send packets with IDs that have random suffixes send(1010 0110,data) S1 1010 0010 A S1 S2 1010 0101 S2 1010 1010 S3 S3 send(1010 1110,data) 1010 1101 S4 S4 B

Example 4: Proximity Suffixes of trigger and packet IDs encode the server and client locations S2 S3 S1 send(1000 0011,data) 1000 1010 S2 1000 1101 S3 1000 0010 S1

Outline Implementation Examples Security Applications Protection against DoS attacks Routing as a service Service composition platform

Applications: Protecting Against DoS Problem scenario: attacker floods the incoming link of the victim Solution: stop attacking traffic before it arrives at the incoming link Today: call the ISP to stop the traffic, and hope for the best! Our approach: give end-host control on what packets to receive Enable end-hosts to stop the attacks in the network

Why End-Hosts (and not Network)? End-hosts can better react to an attack Aware of semantics of traffic they receive Know what traffic they want to protect End-hosts may be in a better position to detect an attack Flash-crowd vs. DoS

Some Useful Defenses White-listing: avoid receiving packets on arbitrary ports Traffic isolation: Contain the traffic of an application under attack Protect the traffic of established connections Throttling new connections: control the rate at which new connections are opened (per sender)

IDP – public trigger IDS, IDR – private triggers 1. White-listing Packets not addressed to open ports are dropped in the network Create a public trigger for each port in the white list Allocate a private trigger for each new connection IDS S Sender (S) Receiver (R) [IDR] IDR R data IDP [IDS] IDP – public trigger IDS, IDR – private triggers

2. Traffic Isolation Drop triggers being flooded without affecting other triggers Protect ongoing connections from new connection requests Protect a service from an attack on another services Victim (V) Attacker (A) Legitimate client (C) ID2 V ID1 Transaction server Web server

2. Traffic Isolation (cont’d) Drop triggers being flooded without affecting other triggers Protect ongoing connections from new connection requests Protect a service from an attack on another services Victim (V) Attacker (A) Legitimate client (C) ID1 V Transaction server Web server Traffic of transaction server protected from attack on web server

3. Throttling New Connections Redirect new connection requests to a gatekeeper Gatekeeper has more resources than victim Can be provided as a 3rd party service Server (S) Client (C) IDC C X S puzzle puzzle’s solution Gatekeeper (A) IDP A

Service Composition Platform Goal: allow third-parties and end-hosts to easily insert new functionality on data path E.g., firewalls, NATs, caching, transcoding, spam filtering, intrusion detection, etc.. Why i3? Make middle-boxes part of the architecture Allow end-hosts/third-parties to explicitly route through middle-boxes

Example Use Bro system to provide intrusion detection for end-hosts that desire so Bro (middle-box) M (idM:idBA, data) (idBA, data) (idAB, data) (idM:idAB, data) idM M server B idBA B client A idAB A i3

Design Principles Give hosts control on routing A trigger is like an entry in a routing table! Flexibility, customization End-hosts can Source route Set-up acyclic communication graphs Route packets through desired service points Stop flows in infrastructure … Implement data forwarding in infrastructure Efficiency, scalability

Design Principles (cont’d) Infrastructure Host Internet & Infrastructure overlays Data plane Control plane p2p & End-host overlays Data plane Control plane i3 Control plane Data plane

Conclusions Indirection – key technique to implement basic communication abstractions Multicast, Anycast, Mobility, … I3 Advocates for building an efficient Indirection Layer on top of IP Explore the implications of changing the communication abstraction; already done in other fields Direct addressable vs. associative memories Point-to-point communication vs. Tuple space (in Distributed systems)

Requirements for Today/Tomorrow’s Internet? Some key requirements (“-ities”) Availability and reliability “Always on”, fault-tolerant, fast recovery from failures, … Quality-of-service (QoS) for applications fast response time, adequate quality for VoIP, IPTV, etc. Scalability millions or more of users, devices, … Mobility untethered access, mobile users, devices, … Security (and Privacy?) protect against malicious attacks, accountability of user actions? Manageability configure, operate and manage networks trouble-shooting network problems Flexibility, Extensibility, Evolvability, ……? ease of new service creation and deployment? evolvable to meet future needs?

Key Issues, Challenges, Solutions … A More Network-Centric View New Naming/Addressing? Separating “identifiers” and “locators” to better support mobility “semantic-free” flat id space ? Data centric? Role of “search” on naming, etc. Scalable and Robust Routing Better and more adaptive to failures, and other network events Also better support for network management, security, … how to perform routing on “flat id” space? Or shall we decouple routing from “naming” or “addressing” ? Manageability “Centralized” approach …? Security (and Privacy?) More “accountable” networks, e.g., through “naming,” or id management?

Key Issues, Challenges, Solutions … Applications and Technology are Dual Drivers ! More devices are connected, novel technologies, disruptive new applications/services Google, and its impact of how we access Internet today social networking: Facebook, MySpace, … iPod/iTune, Skype, BitTorrent, P2P video streaming, YouTube, Hulu.com, Kindle and Amazon, Ebay, … smart phones, etc., “third screen” “Cloud computing”, data centers, and “software as services” Flexibility, Evolvability, and Economic Viability of Network Architectures! It’s “service”, stupid! But is network a (shared) “utility”, “commodity”, or “service” ? “Networks” as services (e.g., VPNs), network security as services, … Network Virtualization and Virtualized Network Architectures User/application “customize-able” network services? Ultimately, networks should be “invisible” !